Inspiration
RetroMate was inspired by the nostalgic appeal of arcade aesthetics combined with the power of Generative AI. The Year of AI theme challenged us to combine retro vibes with cutting-edge technology, creating an “RetroMate” that feels both familiar and innovative.
What it does
RetroMate offers users a safe, interactive interface where they can discuss their thoughts with a character-inspired AI Assistant, “Mario.” Powered by OpenAi, RetroMate uses well-crafted prompt engineering for character-specific responses. With Text-to-Speech for realistic dialogue and InterSystems IRIS Vector Search to retain memory of user interactions, RetroMate creates a holistic conversational experience with a blend of retro arcade aesthetics and AI functionality.
How we built it
Ideation & Brainstorming - We brainstormed and finalized the idea of a retro arcade therapist. OpenAI Integration - Implemented GPT for personalized responses. Text-to-Speech - Text-to-Speech API was added to give our character a voice. IRIS Vector Search - Integrated to retain user interaction memory for more meaningful follow-ups. Front-End Design - Designed the interface using arcade-inspired visuals, for familiarity using Html,Css,Js. Backend - Flask was used as a framework to integrate the whole app Authentication - Set up a basic user login screen, aligning with the retro design for simplicity and privacy.
Challenges we ran into
TTS Model Training and GPU limitations.
Accomplishments that we're proud of
Successfully integrating multiple services like Open AI, Text-to-Speech, and IRIS Vector Search.
What we learned
Technical Skills: Improved skills in integrating diverse AI tools (OpenAI, TTS, IRIS Data Platform ,Vector search). Teamwork: Each phase required coordination, helping us improve our project management and collaboration. User-Centric Design: Learned the importance of interface design that balances functionality with visual appeal for accessibility.
What's next for RETROMATE
Enhanced Personalization: Integrating more characters with unique personalities for varied user interactions. Mobile App Development: Expanding RetroMate to mobile platforms. Improved AI Memory: Enhancing IRIS’s vector search to retain longer-term user memories, enabling a more continuous user experience. Speech-to-Text: Adding Speech-to-Text capabilities for smoother, voice-driven conversations.
Log in or sign up for Devpost to join the conversation.